This question is related to application of rule engines in general, not specifically to drools. Typically, rule engines provide an API to insert facts to working-memory and then executes the rules. In most of the applications, this data originates from a data-source. It is the application responsibility to retrieve the relevant data and to insert in to the engine. If for some reason the query for data is too selective, the engine may end-up with wrong conclusions. So, it is not enough to maintain the business-logic as rules only, there is a need to maintain also the query. What are the common patterns used by applications? * retrieve all data regarding a subject (e.g., in the case of a clinical application - the complete set of patient data). this may cause the collection of large quantities of data (we know that the 'data-gathering' step is much longer than processing it in a rule-engine) * to retrieve only relevant data. Here we will try to restrict the data retrieval only to data that may affect the rules. For example, in the case of a clinical application, only the hemoglobin laboratory results from the last 4 months. The issue here is that query is affected from the rules. We'd like to automatically generate/ validate the data query according to the rules logic. I imagine is an analyzer that receives as input a knowledge-base and returns the greatest-common-query (like GCD). The returned query can be expressed in OCL for example. Does drolls provides a meta-data API to analyze the rule WHEN part? Thanks, David -- View this message in context: http://www.nabble.com/rule-engines-and-data-source-tp25281811p25281811.html Sent from the drools - user mailing list archive at Nabble.com.